X has told Ofcom, Britain’s communications regulator, that it is prepared to review reports of suspected illegal content involving terrorism and hate speech in the United Kingdom much faster in future. On average, assessments are to take place within 24 hours. At least 85% of reports are to be dealt with within 48 hours.
The platform has also undertaken to provide the British regulator with quarterly reports. Ofcom has thus gained a permanent means of monitoring compliance.
The conflict between X and Ofcom shows where platform regulation is heading in practice. It is not merely about the removal of illegal content. In a state governed by the rule of law, a court should decide what is unlawful and what is not. Here, those decisions are made by employees of a tech company whose legal competence may at least be open to question and who, in order to avoid threatened fines for their employer, may prefer to remove too much rather than too little.
Yet the issue goes much further. It also concerns the conditions under which the limits of permissible speech on online platforms are decided.
Sober and Technocratic
The requirements may look like sober technical benchmarks. Their significance, however, goes far beyond that.
Once a platform is judged by how quickly it reacts and how many reports it processes, the incentives change. Leaving questionable content online for too long can create trouble with the regulator. The rapid removal of lawful content, by contrast, often remains invisible and is therefore not punished.
The result is a climate in which speed and caution carry more weight than careful case-by-case assessments of what constitutes legitimate expression.
Protecting the public from terrorist propaganda and calls for violence is a legitimate aim. No free society governed by the rule of law is obliged to tolerate such content.
At the same time, the distinction is crucial. Illegal content is not the same as hurtful, polemical or tasteless opinion. With artificial concepts such as hate speech in particular, much depends on how the term is understood in law and how carefully individual cases are examined.
What Is Truth?
Ofcom justifies the pressure by pointing to the continued presence of illegal content on major platforms. The regulator cites numerous reports involving terrorist material and illegal hate speech. But that is a statement by an executive authority, not a court ruling.
The decisive question is therefore how the procedures are applied. When reporting tools, platform reviews and regulatory audits work together, an opaque machinery of content control emerges. It does not decide abstractly on freedom of expression. Instead, its effects accumulate through countless individual reports, reviews and removal decisions.
The role of external organizations that helped Ofcom gather evidence also deserves attention. They include the Center for Countering Digital Hate. The NGO describes itself as an actor in the fight against online hate and disinformation. Critics accuse it of pursuing political goals of its own and exerting pressure on platforms.
To illustrate the problem, the terms need to be clarified. Hate is a negatively loaded emotion. It is no more illegal to feel it than to express it. Threats of violence arising from hate may indeed be illegal. A clear distinction is indispensable here. Disinformation, too, is an elastic term.
Who decides, then, what is true and what is disinformation? Anyone who described the SSP5-8.5 climate scenario as implausible only a short time ago risked being accused of spreading fake news. Today, even mainstream climate science treats very high-emissions pathways far more cautiously than before.
What was branded disinformation yesterday may become a legitimate position tomorrow. When external expertise is used to assess a statement, it is therefore essential that outside experts be transparent, plural, clearly separated from regulatory decision-making power and demonstrably free from any agenda of their own. In other words, an opinion, even a minority opinion, must be allowed to stand.
A Turning Point in Corporate Policy
For X, the pledge to Ofcom marks a turning point. The platform had previously been sharply critical of the Online Safety Act and warned of interference with freedom of expression. It is now accepting concrete obligations, targets and reporting duties. The shift shows how effective regulatory pressure can be. Anyone wishing to operate in the United Kingdom must come to terms with Ofcom’s rules.
Supporters see this as an overdue step. Illegal content should be reviewed and removed quickly, especially when it involves terrorism or criminal incitement. From a free-speech perspective, however, the question remains how to prevent lawful expression from disappearing as a precaution. The shorter the deadlines and the stricter the audits, the more important transparent standards, explanations and routes for correction become.
The Online Safety Act is intended to protect users and make platforms more accountable. At the same time, the agreement with X shows that such laws do not operate only through outright bans. Content-reporting mechanisms, deadlines, benchmarks, external tipsters and regulatory reporting duties all add to the pressure on platforms. Heavy fines can pose a real threat to their economic survival.
The Asymmetry of Perception
For users, the decisive asymmetry is clear. Processing times can be measured. Removal rates can be tracked. The consequences for free speech are far harder to capture, if they can be captured at all.
There are no statistics on posts that never appear, or on opinions that users do not risk expressing because they feel uncertain. Experts call this the chilling effect: citizens withdraw from debate because they fear saying the wrong thing.
Freedom, including freedom of speech, dies very slowly.
The following example shows what that can look like in practice. TikTok blocked and removed a campaign video by Reform UK’s home affairs spokesman, Zia Yusuf. It had been reported under the content-reporting system of the Online Safety Act. TikTok cited its rules on “hate speech and hateful behavior”. The video dealt with immigration policy and was posted during the campaign for a by-election. The episode illustrates how content moderation can reach into political campaigning.
The agreement with X also shows that cooperation with the regulator can give a company greater certainty. The result is a powerful incentive to err on the side of removal. Platforms end up caught between authorities, NGOs and economic risk.
Resistance creates uncertainty, while excessive removal, or overblocking, reassures regulators and boardrooms but gradually narrows the public space for free expression.